Institute Homepage
Institute Homepage EN Sign In

Publications

DEPARTMENTS

Emperical Interference

Haptic Intelligence

Modern Magnetic Systems

Perceiving Systems

Physical Intelligence

Robotic Materials

Social Foundations of Computation

Research Groups

Autonomous Vision

Autonomous Learning

Bioinspired Autonomous Miniature Robots

Dynamic Locomotion

Embodied Vision

Human Aspects of Machine Learning

Intelligent Control Systems

Learning and Dynamical Systems

Locomotion in Biorobotic and Somatic Systems

Micro, Nano, and Molecular Systems

Movement Generation and Control

Neural Capture and Synthesis

Physics for Inference and Optimization

Organizational Leadership and Diversity

Probabilistic Learning Group

Topics

Robot Learning

Conference Paper

2022

Autonomous Learning

Robotics

AI

Career

Award

Autonomous Motion Conference Paper Programmable pattern generators Schaal, S., Sternad, D. In 3rd International Conference on Computational Intelligence in Neuroscience, 48-51, Research Triangle Park, NC, Oct. 24-28, October 1998, clmc
This paper explores the idea to create complex human-like arm movements from movement primitives based on nonlinear attractor dynamics. Each degree-of-freedom of an arm is assumed to have two independent abilities to create movement, one through a discrete dynamic system, and one through a rhythmic system. The discrete system creates point-to-point movements based on internal or external target specifications. The rhythmic system can add an additional oscillatory movement relative to the current position of the discrete system. In the present study, we develop appropriate dynamic systems that can realize the above model, motivate the particular choice of the systems from a biological and engineering point of view, and present simulation results of the performance of such movement primitives. Implementation results on a Sarcos Dexterous Arm are discussed.
URL BibTeX
Empirical Inference Article SVMs — a practical consequence of learning theory Schölkopf, B. IEEE Intelligent Systems and their Applications, 13(4):18-21, July 1998
My first exposure to Support Vector Machines came this spring when heard Sue Dumais present impressive results on text categorization using this analysis technique. This issue's collection of essays should help familiarize our readers with this interesting new racehorse in the Machine Learning stable. Bernhard Scholkopf, in an introductory overview, points out that a particular advantage of SVMs over other learning algorithms is that it can be analyzed theoretically using concepts from computational learning theory, and at the same time can achieve good performance when applied to real problems. Examples of these real-world applications are provided by Sue Dumais, who describes the aforementioned text-categorization problem, yielding the best results to date on the Reuters collection, and Edgar Osuna, who presents strong results on application to face detection. Our fourth author, John Platt, gives us a practical guide and a new technique for implementing the algorithm efficiently.
PDF Web DOI BibTeX
Empirical Inference Conference Paper From regularization operators to support vector kernels Smola, A., Schölkopf, B. In Advances in Neural Information Processing Systems, Advances in Neural Information Processing Systems 10, 343-349, (Editors: M Jordan and M Kearns and S Solla), MIT Press, Cambridge, MA, USA, 11th Annual Conference on Neural Information Processing (NIPS 1997), June 1998 PDF Web BibTeX
Empirical Inference Conference Paper Prior knowledge in support vector kernels Schölkopf, B., Simard, P., Smola, A., Vapnik, V. In Advances in Neural Information Processing Systems, Advances in Neural Information Processing Systems 10, 640-646 , (Editors: M Jordan and M Kearns and S Solla ), MIT Press, Cambridge, MA, USA, Eleventh Annual Conference on Neural Information Processing (NIPS 1997), June 1998 PDF Web BibTeX
Empirical Inference Article Learning view graphs for robot navigation Franz, M., Schölkopf, B., Mallot, H., Bülthoff, H. A. Autonomous Robots, 5(1):111-125, March 1998
We present a purely vision-based scheme for learning a topological representation of an open environment. The system represents selected places by local views of the surrounding scene, and finds traversable paths between them. The set of recorded views and their connections are combined into a graph model of the environment. To navigate between views connected in the graph, we employ a homing strategy inspired by findings of insect ethology. In robot experiments, we demonstrate that complex visual exploration and navigation tasks can thus be performed without using metric information.
PDF PDF DOI BibTeX
Physical Intelligence Conference Paper 2D micro particle assembly using atomic force microscope Sitti, M., Hirahara, K., Hashimoto, H. In Micromechatronics and Human Science, 1998. MHS’98. Proceedings of the 1998 International Symposium on, 143-148, 1998 BibTeX
Autonomous Motion Conference Paper Biomimetic gaze stabilization based on a study of the vestibulocerebellum Shibata, T., Schaal, S. In European Workshop on Learning Robots, 84-94, Edinburgh, UK, 1998, clmc
Accurate oculomotor control is one of the essential pre-requisites for successful visuomotor coordination. In this paper, we suggest a biologically inspired control system for learning gaze stabilization with a biomimetic robotic oculomotor system. In a stepwise fashion, we develop a control circuit for the vestibulo-ocular reflex (VOR) and the opto-kinetic response (OKR), and add a nonlinear learning network to allow adaptivity. We discuss the parallels and differences of our system with biological oculomotor control and suggest solutions how to deal with nonlinearities and time delays in the control system. In simulation and actual robot studies, we demonstrate that our system can learn gaze stabilization in real time in only a few seconds with high final accuracy.
URL BibTeX
Autonomous Motion Article Constructive incremental learning from only local information Schaal, S., Atkeson, C. G. Neural Computation, 10(8):2047-2084, 1998, clmc
We introduce a constructive, incremental learning system for regression problems that models data by means of spatially localized linear models. In contrast to other approaches, the size and shape of the receptive field of each locally linear model as well as the parameters of the locally linear model itself are learned independently, i.e., without the need for competition or any other kind of communication. Independent learning is accomplished by incrementally minimizing a weighted local cross validation error. As a result, we obtain a learning system that can allocate resources as needed while dealing with the bias-variance dilemma in a principled way. The spatial localization of the linear models increases robustness towards negative interference. Our learning system can be interpreted as a nonparametric adaptive bandwidth smoother, as a mixture of experts where the experts are trained in isolation, and as a learning system which profits from combining independent expert knowledge on the same problem. This paper illustrates the potential learning capabilities of purely local learning and offers an interesting and powerful approach to learning with receptive fields. 
URL BibTeX
Micro, Nano, and Molecular Systems Article Linear electro-optic effect in optically active liquids Buckingham, A., Fischer, P. CHEMICAL PHYSICS LETTERS, 297(3-4):239-246, 1998
A linear effect of an electrostatic field F on the intensity of sum- and difference-frequency generation in a chiral liquid is predicted. It arises in the electric dipole approximation. The effect changes sign with the enantiomer and on reversing the direction of the electrostatic field. The sum-frequency generator chi(alpha beta gamma)((2)) (-omega(3);omega(1),omega(2)), where omega(3) = omega(1) + omega(2), and the electric field-induced sum-frequency generator chi(alpha beta gamma delta)((3))(-omega(3);omega(1),omega(2),0)F-delta interfere and their contributions to the scattering power can be distinguished. Encouraging predictions are given for a typical experimental arrangement. (C) 1998 Elsevier Science B.V. All rights reserved.
DOI BibTeX
Autonomous Motion Article Local adaptive subspace regression Vijayakumar, S., Schaal, S. Neural Processing Letters, 7(3):139-149, 1998, clmc
Incremental learning of sensorimotor transformations in high dimensional spaces is one of the basic prerequisites for the success of autonomous robot devices as well as biological movement systems. So far, due to sparsity of data in high dimensional spaces, learning in such settings requires a significant amount of prior knowledge about the learning task, usually provided by a human expert. In this paper we suggest a partial revision of the view. Based on empirical studies, we observed that, despite being globally high dimensional and sparse, data distributions from physical movement systems are locally low dimensional and dense. Under this assumption, we derive a learning algorithm, Locally Adaptive Subspace Regression, that exploits this property by combining a dynamically growing local dimensionality reduction technique  as a preprocessing step with a nonparametric learning technique, locally weighted regression, that also learns the region of validity of the regression. The usefulness of the algorithm and the validity of its assumptions are illustrated for a synthetic data set, and for data of the inverse dynamics of human arm movements and an actual 7 degree-of-freedom anthropomorphic robot arm. 
URL BibTeX
Autonomous Motion Conference Paper Local dimensionality reduction Schaal, S., Vijayakumar, S., Atkeson, C. G. In Advances in Neural Information Processing Systems 10, 633-639, (Editors: Jordan, M. I.;Kearns, M. J.;Solla, S. A.), MIT Press, Cambridge, MA, 1998, clmc
If globally high dimensional data has locally only low dimensional distributions, it is advantageous to perform a local dimensionality reduction before further processing the data. In this paper we examine several techniques for local dimensionality reduction in the context of locally weighted linear regression. As possible candidates, we derive local versions of factor analysis regression, principle component regression, principle component regression on joint distributions, and partial least squares regression. After outlining the statistical bases of these methods, we perform Monte Carlo simulations to evaluate their robustness with respect to violations of their statistical assumptions. One surprising outcome is that locally weighted partial least squares regression offers the best average results, thus outperforming even factor analysis, the theoretically most appealing of our candidate techniques.
URL BibTeX
Physical Intelligence Conference Paper Macro to nano tele-manipulation through nanoelectromechanical systems Sitti, M., Hashimoto, H. In Industrial Electronics Society, 1998. IECON’98. Proceedings of the 24th Annual Conference of the IEEE, 1:98-103, 1998 BibTeX
Micro, Nano, and Molecular Systems Article Monolayers of hexadecyltrimethylammonium p-tosylate at the air-water interface. 1. Sum-frequency spectroscopy Bell, G., Li, Z., Bain, C., Fischer, P., Duffy, D. JOURNAL OF PHYSICAL CHEMISTRY B, 102(47):9461-9472, 1998
Sum-frequency vibrational spectroscopy has been used to determine the structure of monolayers of the cationic surfactant, hexadecyltrimethylammonium p-tosylate (C(16)TA(+)Ts(-)), at the surface of water. Selective deuteration of the cation or the anion allowed the separate detection of sum-frequency spectra of the surfactant and of counterions that are bound to the monolayer. The p-tosylate ions an oriented with their methyl groups pointing away from the aqueous subphase and with the C-2 axis tilted, on average, by 30-40 degrees from the surface normal. The vibrational spectra of C(16)TA(+) indicate that the number of gauche defects in the monolayer does not change dramatically when bromide counterions are replaced by p-tosylate. The ends of the hydrocarbon chains of C16TA+ are, however, tilted much further from the surface normal in the presence of p-tosylate than in the presence of bromide. A quantitative analysis of the sum-frequency spectra requires a knowledge of the molecular hyperpolarizability tensor: the role of ab initio calculations and Raman spectroscopy in determining the components of this tensor is discussed.
DOI BibTeX
Physical Intelligence Conference Paper Nano tele-manipulation using virtual reality interface Sitti, M., Horiguchi, S., Hashimoto, H. In Industrial Electronics, 1998. Proceedings. ISIE’98. IEEE International Symposium on, 1:171-176, 1998 BibTeX
Autonomous Motion Conference Paper Robust local learning in high dimensional spaces Vijayakumar, S., Schaal, S. In 5th Joint Symposium on Neural Computation, 186-193, Institute for Neural Computation, University of California, San Diego, San Diego, CA, 1998, clmc
Incremental learning of sensorimotor transformations in high dimensional spaces is one of the basic prerequisites for the success of autonomous robot devices as well as biological movement systems. So far, due to sparsity of data in high dimensional spaces, learning in such settings requires a significant amount of prior knowledge about the learning task, usually provided by a human expert. In this paper, we suggest a partial revision of this view. Based on empirical studies, we observed that, despite being globally high dimensional and sparse, data distributions from physical movement systems are locally low dimensional and dense. Under this assumption, we derive a learning algorithm, Locally Adaptive Subspace Regression, that exploits this property by combining a dynamically growing local dimensionality reduction technique as a preprocessing step with a nonparametric learning technique, locally weighted regression, that also learns the region of validity of the regression. The usefulness of the algorithm and the validity of its assumptions are illustrated for a synthetic data set, and for data of the inverse dynamics of human arm movements and an actual 7 degree-of-freedom anthropomorphic robot arm.
BibTeX
Micro, Nano, and Molecular Systems Article Surface second-order nonlinear optical activity Fischer, P., Buckingham, A. JOURNAL OF THE OPTICAL SOCIETY OF AMERICA B-OPTICAL PHYSICS, 15(12):2951-2957, 1998
Following the recent observation of a large second-harmonic intensity difference from a monolayer of chiral molecules with left and right circularly polarized light, the scattering theory is generalized and extended to predict linear and circular intensity differences for the more Versatile sum-frequency spectroscopy. Estimates indicate that intensity differences should be detectable for a typical experimental arrangement. The second-order nonlinear surface susceptibility tensor is given for different surface point groups in the electric dipole approximation; it is shown that nonlinear optical activity phenomena unambiguously probe molecular chirality only for molecular monolayers that are symmetric about the normal. Other surface symmetries can give rise to intensity differences from monolayers composed of achiral molecules. A water surface is predicted to show Linear and nonlinear optical activity in the presence of an electric field parallel to the surface. (C) 1998 Optical Society of America {[}S0740-3224(98)01311-3] OCIS codes: 190.0190, 190.4350, 240.6490.
DOI BibTeX
Physical Intelligence Article Tele-nanorobotics using an atomic force microscope as a nanorobot and sensor Sitti, M., Hashimoto, H. Advanced Robotics, 13(4):417-436, Taylor & Francis, 1998 BibTeX
Physical Intelligence Conference Paper Tele-nanorobotics using atomic force microscope Sitti, M., Hashimoto, H. In Intelligent Robots and Systems, 1998. Proceedings., 1998 IEEE/RSJ International Conference on, 3:1739-1746, 1998 BibTeX
Autonomous Motion Conference Paper Towards biomimetic vision Shibata, T., Schaal, S. In International Conference on Intelligence Robots and Systems, 872-879, Victoria, Canada, 1998, clmc
Oculomotor control is the foundation of most biological visual systems, as well as an important component in the entire perceptual-motor system. We review some of the most basic principles of biological oculomotor systems, and explore their usefulness from both the biological and computational point of view. As an example of biomimetic oculomotor control, we present the state of our implementations and experimental results using the vestibulo-ocular-reflex and opto-kinetic-reflex paradigm
URL BibTeX
Micro, Nano, and Molecular Systems Article Ultraviolet resonance Raman study of drug binding in dihydrofolate reductase, gyrase, and catechol O-methyltransferase Couling, V., Fischer, P., Klenerman, D., Huber, W. BIOPHYSICAL JOURNAL, 75(2):1097-1106, 1998
This paper presents a study of the use of ultraviolet resonance Raman (UVRR) spectroscopic methods as a means of elucidating aspects of drug-protein interactions. Some of the RR vibrational bands of the aromatic amino acids tyrosine and tryptophan are sensitive to the microenvironment, and the use of UV excitation radiation allows selective enhancement of the spectral features of the aromatic amino acids, enabling observation specifically of their change in microenvironment upon drug binding. The three drug-protein systems investigated in this study are dihydrofolate reductase with its inhibitor trimethoprim, gyrase with novobiocin, and catechol O-methyltransferase with dinitrocatechol. It is demonstrated that UVRR spectroscopy has adequate sensitivity to be a useful means of detecting drug-protein interactions in those systems for which the electronic absorption of the aromatic amino acids changes because of hydrogen bonding and/or possible dipole-dipole and dipole-polarizability interactions with the ligand.
DOI BibTeX